Questions re metadata tags & filters

[chrisdemaagd]

We're currently experiencing challenges in implementing faceted search w/i the doc portal (Zoomin) my company uses. The keywords we have configured are not appearing in the facets on the live production portal, which appears to be related to a metadata or publishing chain issue. (This situation seems to stem from our migration to GitHub/Oxygen, which resulted in some original metadata being lost.) The core issue is that the keywords we've set up for faceted search aren't populating the facets on the doc portal. We've been working to correct this by re-implementing the structured keywords in our DITA maps, which are intended to link up with the facets on the doc portal. Despite placing the keywords and publishing the content with them, the facets remain empty. We're trying to pinpoint where in the Oxygen --> Zoomin transformation chain the metadata is being lost or misconfigured.

Below are key details about our setup and specific questions:

Keyword Location & Structure
We now use two types of keywords:

• Structured keywords: Located in the DITA maps (intended for facets)

• Unstructured keywords: Embedded at the topic level within the <prolog>

Questions:
1. Where should the facet keywords reside? Are they expected exclusively in the DITA map, or should structured keywords also appear at the topic level, inside <prolog>?
2. Is our present approach for placing structured keywords in the DITA map (e.g., via <topicmeta> or designated attributes) correct?
Transformation Scenario & XSLT

Our publishing workflow is as follows: Oxygen → Zoomin Transformation Scenario → Documentation Portal.

Questions:
1. Does the transformation scenario require adjustments to ensure that it correctly extracts and passes the relevant DITA map or topic-level metadata used for facets?
2. Is it necessary to explicitly configure arguments in the transformation scenario to manage or map facet keywords?
3. Is the problem likely with our custom XSLT? Are there specific XSLT files responsible for mapping this DITA map metadata? Is there an XSLT customization or hook that could be misconfigured during the transformation process? Should we update the scenario to address the XSLT files?

Any insights, particularly from those with experience in Oxygen metadata setup and faceted filter configuration, would be greatly valued.
Thank you!

[Radu]

Hi,
We (the Oxygen team) do not know how the Zoomin publishing works or what features it has.
This seems like the kind of question you should ask Zoomin as they control the publishing part.
Maybe you can create a small sample DITA Map exhibiting how you set up the keyword using Oxygen, then send them the sample DITA Map and ask for help to structure the keywords in order to obtain the proper publishing format.
Regards,
Radu